response property of cation

From NWChem

Viewed 2502 times, With a total of 3 Posts
Jump to: navigation, search

Just Got Here
Threads 2
Posts 4
Hello

I want to compute static polartizablity of Na+
I am using development version with following info
   nwchem branch   = Development
nwchem revision = 24322
ga revision = 10386

I have used following input file


echo

start Na

title "Na"

memory stack 800 mb heap 100 mb global 1500 mb noverify

permanent_dir .
scratch_dir /tmp

geometry
symmetry d2h
Na 0.000000 0.000000 0.000000
end

CHARGE 1

basis "adz" spherical noprint
 * library aug-cc-pvdz
end

basis "dadz" spherical noprint
 * library d-aug-cc-pvdz
end

basis "datz" spherical noprint
 * library d-aug-cc-pvtz
end

set "ao basis" "adz"


set int:acc_std 1d-20

scf
 rhf
singlet
semidirect memsize 100000000 filesize 1000000
thresh 1e-10
tol2e 1e-16
maxiter 100000
print "final evals"
end

tce
  1. freeze atomic
 ccsd
thresh 1e-10
maxiter 10000
io ga
2eorb
2emet 12
tilesize 32
attilesize 32
diis 8
diis2 8
diis3 8
lshift 0.1 # CC-T
lshiftl 0.1 # CC-Lambda
lshift2 0.4 0.2 # CC-LR
lshift3 0.5 0.2 # CC-IR
end


set tce:lineresp T
set tce:afreq 0.0
set tce:respaxis T F F
task tce energy


but the job terminate with error.


Integral file          = /tmp/Na.aoints.00
Record size in doubles = 65536 No. of integs per rec = 43688
Max. records in memory = 1525 Max. records in file = 14
No. of bits per label = 8 No. of bits per value = 64


File balance: exchanges= 0 moved= 0 time= 0.0


Fock matrix recomputed
1-e file size = 141
1-e file name = /tmp/Na.f1
Cpu & wall time / sec 0.3 0.5
4-electron integrals stored in orbital form
1:Floating Point Exception error, status=: 8
(rank:1 hostname:thnec169 pid:18855):ARMCI DASSERT fail. ../../ga-5-2/armci/src/common/signaltrap.c:SigFpeHandler():249 cond:0
rank 1 in job 1 thnec169_45791 caused collective abort of all ranks
 exit status of rank 1: killed by signal 9



Is this a bug or some error in my input file?
I tried to tune so many parameter but i get same error.

Regards
Vivekanand

Forum Vet
Threads 9
Posts 1395
Switch to 2emet 13
Vivekanand,
The failure is caused by the algorithm used when the 2emet keyword is set equal to 12.
Please set 2emet equal 13 and you calculation will run successfully.
E.g.:
2emet 13

Cheers, Edo

PS: You might want to modify the tce:respaxis line, too, to have all three components calculated
(e.g. set tce:respaxis T T T), since when only one is calculated, you might get a failure.
Edited On 2:47:51 PM PDT - Fri, Aug 16th 2013 by Edoapra

Just Got Here
Threads 2
Posts 4
Many Thank Edo
by adding necessary keyword mentioned in last thread. i.e.
2emet 13
&
set tce:lineresp T
set tce:afreq 0.0
set tce:respaxis T T T

Now the calculation runs suscessfully for static polarizability.

My original interest was to compute dynamic polarizability at imaginary frequecy (20 point gauss-legendre grid)
therefore, i change the "2emet 13" as done above and change TCE response block to


set tce:lineresp F
set tce:imagresp T
set tce:ifreqauto F
set tce:ifreq 0.0 0.0392901 0.118358 0.198912 0.282029 0.368919 0.461006 0.560027 0.668179 0.788336 0.92439
1.08179 1.26849 1.49661 1.78563 2.16917 2.71062 3.54573 5.02734 8.44896 25.4517




XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX---------------------------
Query 1: "YOU WILL GET WRONG RESULTS " what is the meaning of this line in the output?


Singles contributions

Doubles contributions

CCSD Lambda iterations
---------------------------------------------
Iter Residuum Cpu Wall
---------------------------------------------
1 3.5217657001461 0.1 0.1
2 0.0066280791423 0.1 0.1
3 0.0002076259247 0.1 0.1
4 0.0000063048554 0.1 0.1
5 0.0000006893744 0.1 0.1
6 0.0000000515447 0.1 0.1
7 0.0000000051074 0.1 0.1
8 0.0000000004770 0.1 0.1
MICROCYCLE DIIS UPDATE: 8 8
9 0.0000000000004 0.1 0.1
---------------------------------------------
Iterations converged

Singles contributions

Doubles contributions

CCSD dipole moments / hartree & Debye
------------------------------------
X 0.0000000 0.0000000
Y 0.0000000 0.0000000
Z 0.0000000 0.0000000
Total 0.0000000 0.0000000
------------------------------------
YOU WILL GET WRONG RESULTS

Casimir-Polder integration points are:
ifreq( 1) = 0.00000000
ifreq( 2) = 0.03929010
ifreq( 3) = 0.11835800
ifreq( 4) = 0.19891200
ifreq( 5) = 0.28202900
ifreq( 6) = 0.36891900
ifreq( 7) = 0.46100600
ifreq( 8) = 0.56002700
ifreq( 9) = 0.66817900
......




Query 2: Now the calculation run and compute dynamic polarizability tensor upto "25.4517"
However instested of usual "normal termination" message.
I get following message
.......(tailed error output)
CCSD-IR (imaginary component) - linear iterations
---------------------------------------------
Iter Residuum Cpu Wall
---------------------------------------------
1 0.0717594666302 0.1 0.1
2 0.0245121765210 0.1 0.1
3 0.0077031963582 0.1 0.1
4 0.0025309268597 0.1 0.1
5 0.0009248494453 0.1 0.1
6 0.0003634115130 0.1 0.1
7 0.0001476377461 0.1 0.1
8 0.0000608007544 0.1 0.1
MICROCYCLE DIIS UPDATE: 8 8
9 0.0000002558741 0.1 0.1
10 0.0000000471362 0.1 0.1
11 0.0000000117436 0.1 0.1
12 0.0000000033748 0.1 0.1
13 0.0000000010547 0.1 0.1
14 0.0000000003583 0.1 0.1
15 0.0000000001296 0.1 0.1
16 0.0000000000492 0.1 0.1
---------------------------------------------
Iterations converged

CCSD Imaginary Response polarizability / au
Frequency = 25.4517000 / au
Wavelength = 1.7901890 / nm
-------------------------------------------------------|-----------------------------------------------
atomic units (bohr^3) | angstroms^3
X Y Z | X Y Z
-------------------------------------------------------|-----------------------------------------------
X 0.0093258 0.0000000 0.0000000 | 0.0013819 0.0000000 0.0000000
Y 0.0000000 0.0093258 0.0000000 | 0.0000000 0.0013819 0.0000000
Z 0.0000000 0.0000000 0.0093258 | 0.0000000 0.0000000 0.0013819
-------------------------------------------------------|-----------------------------------------------
Eigs = 0.0093258 0.0093258 0.0093258 | 0.0013819 0.0013819 0.0013819
Isotropic = 0.0093258 | 0.0013819
Anisotropic = 0.0000000 | 0.0000000
-------------------------------------------------------|-----------------------------------------------
Cpu & wall time / sec 1.6 3.1
------------------------------------------------------------------------
MANUAL STOP 0
------------------------------------------------------------------------
This type of error is most commonly associated with calculations not reaching
convergence criteria
------------------------------------------------------------------------
For more information see the NWChem manual at
http://www.nwchem-sw.org/index.php/NWChem_Documentation


For further details see manual section:




0:0:MANUAL STOP:: -1
(rank:0 hostname:thnec177 pid:24477):ARMCI DASSERT fail. ../../ga-5-2/armci/src/common/armci.c:ARMCI_Error():208 cond:0
MANUAL STOP        0
------------------------------------------------------------------------
MANUAL STOP 0
------------------------------------------------------------------------
MANUAL STOP 0
------------------------------------------------------------------------

Are these error affect the computed tensor components ?

Regards
Vivekanand Gobre

Gets Around
Threads 3
Posts 74
YOU WILL GET WRONG RESULTS
The statement "YOU WILL GET WRONG RESULTS" is meant to warn the user as to the unverified nature of the CCSD imaginary response code. I tried very hard to make it work and rewrote the code half a dozen times but kept getting inconsistencies w.r.t. previous ones. There is no reference work to which I can compare directly.

It is entirely possible that the code is correct but I won't certify that and you should not publish results claiming you've done CCSD C6 in NWChem unless you have an exact numerical verification against an independent implementation.

That the C6 code is undocumented in the user manual should indicate that it is not to be used. The "YOU WILL GET WRONG RESULTS" warning further amplifies that point.

If you want C6 coefficients, I recommend you use the Cauchy moment method based upon the polarizability for real frequencies. It is documented in the literature but not implemented in NWChem. You can write Python or Matlab or ... though since it is not compute-intensive to solve these equations.


Forum >> NWChem's corner >> Running NWChem



Who's here now Members 0 Guests 0 Bots/Crawler 1


AWC's: 2.5.10 MediaWiki - Stand Alone Forum Extension
Forum theme style by: AWC